[00:00:00] Guta: I do think there are a lot of proposed solutions to all these measurement problems. And I think it's about philosophy. In the end, you need to understand how those people think cause algorithms, they do have opinions The next post that you see on your feed. It's opinionated. [00:00:16] so we do make decisions every day about the algorithm and our decisions are made from a company that believes in putting branding into the equation, and creating a unified system to marketing so you can manage, having a lower cost over time and a more solid baseline in organic channels. [00:00:37] ​ [00:01:04] In This Episode --- [00:01:04] Phil: What's up everyone? Today we have the pleasure of sitting down with Guta Tolmasquim, founder and CEO at Purple Metrics. After spending almost a decade at top branding agencies in Brazil, gutta founded and ran her own agency where she rebranded over 200 plus startups. She sold that agency after five years to build Purple Metrics, a marketing attribution software with an algorithm that includes brand impact In this episode, weeks. [00:01:28] Or the tangible outcomes of brand, like purchase intent and memory structures, how to build trust with measurement solutions, how to convince CFOs to fund brand marketing and creating attribution algorithms that respect brand complexity. All that, and a bunch more stuff after. Super quick word from one of our awesome partners. [00:01:47] ​ [00:02:41] Phil: Guta, thank you so much for your time today. Really excited to chat. [00:02:45] Guta: Thanks for having me. I'm very happy to be here. [00:02:48] 1. How Brand Measurement Connects to Revenue in Startups --- [00:02:48] Phil: We spent a lot of time on the show talking about, uh, vendors with consultants about measurements and, and attribution. And many of the experts and products gravitate towards performance [00:03:00] channel and like paid media and the company you're running Purple was built with this like brand first approach to measurement and attribution. [00:03:08] Maybe start by like bringing us back to the first moments when you realize there was a better way to measure brand. [00:03:15] Guta: Okay. Uh, so it's a, a great start, a great opening question. I actually recently, I, I did this talk to 5,000 marketers and I went into the branding history. I went into researching it, and what I realized is that at the beginning, like. 1920s, 1930s, when brands started being added to the marketing departments, they were actually part of the sales team of marketing and perfor performance, as you would call it now. And then at some, some point in history, around the sixties, the seventies, we started splitting branding from sales. [00:04:00] And then we had that, those nineties with those big brands, you know, Pepsi and Victoria Secret and Playboy, and all of that. Those big branding, big brand building time periods in history. And then digital media came on and we all started with their metrics, you know, and tracking everything every click. So it got split more and more, got more separated over the course of the last. 20, 30 years. And I had a different story. So I started in growth marketing in performance media. And then I was like, I don't wanna work with that outta respect and on all fans, but it's not my thing. And I went into branding and I worked, like I had a almost 10 years experience with branding and advertising with enterprises, you know? And I was like, [00:05:00] it's too conceptual. Why don't we like, so I always had, then I went into working with branding for startups and those guys are crazy, crazy over measuring everything and data driven everything. So I always had this branding brand oriented towards metrics career, which is very unusual. And I was like, why can't we measure it? [00:05:27] And people got asking me that over and over again. But I think it was only when I was, uh, at my first company. Seeing the results. 'cause that's, that's something interesting on working with startups. They only have money to do one thing at a time, money and energy. Right? So when they are rebranding or repositioning, they are only doing that. So it's easy to attribute [00:05:52] Phil: Hmm. [00:05:52] Guta: like I did it and sales increase and everything else was, it's like almost an ab tested, uh, mindset [00:06:00] mentality. 'cause you changed one thing and the rest remain the same. And I started to see immediate returns, not only long term and measurable returns 'cause they were measuring everything. So I was like, hmm, maybe there is a way. And then I started looking into it and decided to do something that was a better research tool. So let's do research 'cause that's how everything, everyone measures branding in a more efficient way, more technological way. Let's update brand research. And then I realized, and we did this product, we launched it, we were, uh, on our, our way to product market. [00:06:41] I think we reached product market fit. 'cause we did have, we have companies using it up to today and we haven't updated the products in a year. Uh, but then I was like, if we don't connect to money, to cash to dollars, there's no point. I mean, no one cares if you don't. [00:07:00] No one cares about brand equity. You know, only brand marketers. [00:07:04] I, I do care, but I mean, in general, you can't run a, a business on a, a self-declared purchase intent. [00:07:14] Phil: Yeah, the, the CFO doesn't care. [00:07:17] Guta: It doesn't care. And maybe the CMO shouldn't care as well. I mean, it's an information, but it's not a decisive information that you should be with your entire market marketing strategy towards. Um, so that's. That's, it's, it wasn't like a pivoting moment where when I realized it was like, okay, maybe there's a better way, and then we were like moving towards that. I was moving towards that in the last maybe 10 years, if you think about it, but actively in the last four. [00:07:53] Phil: Gotcha. So now flash forward to the new product. You're focusing a bit more on tying those [00:08:00] brand measurement ideas to actual dollars and, and revenue and like something that we get a lot from this idea that like. How does brand actually drive behaviors? Before we get into the, the dollars there, like [00:08:15] 2. Examples Where Brand Investments Shifted Real Business Outcomes --- [00:08:15] Phil: maybe you can share some examples where brand investments shifted real business outcomes like incremental revenue, loyalty, LTV or pricing power. [00:08:26] What are some of the examples that that come to mind? [00:08:29] Guta: Cool. So we, we did have some examples that were in mindset. Metrics. That's how, how it's called, right? So, um, we had this client and they, they are a bank and they sponsored soccer, a soccer team in the soccer championship. Soccer is huge in Brazil, [00:08:52] Phil: Mm-hmm. [00:08:53] Guta: those of you that haven't, I mean, everyone knows, I mean, that's the only information about Brazil that's worldwide, [00:09:00] uh, known, right? Should no need for explain. But anyways, uh, so it's a great marketing [00:09:06] investment usually, but it's a very hard to track 'cause it's not clickable Taiwan. So we saw that the soccer fans were, uh, more loyal, declared, declared, how do you say that? More loyal to the brand. And this kind of pointed out to them where to look. 'cause then they went into their data. On the soccer fans base and es and they were like, oh, now we have a really dollar connected result, right? To this pure churn. That was one of one insight for us that, oh, why don't we like, why don't we measure the dollar connected thing? But it's usually in, for our clients, it's about knowing what channels convert, [00:10:00] even the organic ones. Because organic, there's no organic. Organic is a myth. It's just something that you don't know what generated it. There is no organic, uh, but then seeing influencers, for example, it's something that appears a lot. And they were like, oh, we had no idea. There was this client of us that they rebranded, so they had the no link name page and the new one. 'cause they put every, every product in under the same brand name. So they had all these multiple LinkedIn profiles and they were just about to discontinue it when they started using purple metrics and they saw it was driving like over 10% of conversion from something that they would discontinue. 'cause no one [00:10:49] clicks on LinkedIn. 'cause LinkedIn doesn't allow links basically, right. Uh, it do, it do allow. It does allow, but like [00:11:00] it doesn't [00:11:00] get any reach. you [00:11:02] Phil: do. Yeah, exactly. [00:11:04] Guta: 20 times less [00:11:05] anyways, so people just go on LinkedIn and then they go on Google, wherever channel that you would attribute to. Uh, so we started seeing these and we started seeing these movements of how the money flows through the company. So all these campaigns are generating, we actually can tell our clients what generates branded search on Google. So you see and, and how long it takes so you can actually see the system working. So this generated that in four weeks and that generated conversion in three weeks. We're like, oh shit, I need to start planning my 4th of July campaign now. [00:11:50] Right. [00:11:53] Phil: When we talk about brand measurement, lots of folks think of terms like brand equity and brand [00:12:00] lift studies, and I feel like that gets a bit of a fluffy reputation. And it's kind of the opposite when we talk about content, folks are trying to force lead targets and MQL targets with content and it, so it's like brand is this like top of funnel thing and we're not really sure how to measure it except for like long-term studies, but content that's way different. [00:12:21] Let, let's put a lead target on that one. [00:12:23] 3. The Tangible Outcomes of Brand: Purchase Intent Considerations, Memory Structures --- [00:12:23] Phil: Talk to us about like the tangible outcomes of brand and the like other outcomes of content that you've kind of researched and investigated. Um, like on your site you talk about like purchase intent considerations, sets and memory structures. I found that really interesting. [00:12:40] Maybe unpack that for us. [00:12:42] Guta: Cool. Um, so branding is something very intangible. 'cause what is, what is it, right? But then when you start understanding what composes the brand perception, you go through content and maybe. [00:13:00] Campaigns, awareness campaigns, people call it in all these different ways, but it's the same or referral or community or events sponsoring things, being on podcasts, thought leadership. [00:13:16] So you have all those actions, marketing actions, and initiatives that we consider branding. Maybe not, maybe still do. And that is actually what drives behavior, right? Because the way that branding works is that it generates some type of memory. It's something that's made to last in your mind. And then by the time you go to market, by the time you are in market for some product, you, your brain downloads it somehow. So your brain goes, oh, that brand, there's it. This is their value prop, and [00:14:00] or this is how they differentiate from others, or something, some attributes, some an attribute that you associate with it and it actually makes performance run faster. Your decision making is faster. That's, that's com that's, there's research on that. So you just pick products more fast when you, faster when you, when you know something about the brand. So those, you need to get it. So you can do two types of communications. You can activate some behavior. So I want fail to click, to download, to subscribe, to experiment, to whatever. So that's one type, one form of communication, which is activating a behavior. And then the other form is, let me tell, feel about my product and my, my setup differentiators and whatever I wanna build [00:15:00] in your memory is. So understanding this balance is very important. And what people do with content is free behavior activating, right. But you can only do it and sometimes you're like, oh, it went viral, but no one clicked. Yeah. 'cause probably a different type of communication. It's something that I'm gonna relate about and then I'm probably gonna act on it if I see something that activates my behavior. So it's always about this balance of two types, two forms of communication. And the, the interesting part of it is that like a Subway, they share the trails, [00:15:43] you know, so you can have the yellow line in green line running in the same tracks and it's so you can have or TikTok and run two types of communications on it. And it, it's actually very [00:16:00] effective as sim data from big texts. And they are doing this in inserting branding creatives into performance campaigns. [00:16:10] And it actually, the entire performance, 'cause it's just reminded you of who I am and then I activate your behavior. Uh, so I think it's a more complex problem than what's generating sales. [00:16:23] Let's just double down on that, you know. [00:16:27] Phil: I love your, your subway analogy because too many people think of it as one or the other. The traditional like brand versus demand gen debates and which dollar do we invest in? And when you talk about this idea of educating the market versus getting people to act and click and buy, when you present those two options to A CFO or the CEO, most of the time they're gonna gravitate towards the, well, yeah, this one's easier to track, and we get dollars out of it. [00:16:56] And the other one traditionally is a lot harder to track [00:17:00] and we don't know if people getting more educated is directly going to lead to more revenue. [00:17:05] 4. How to Convince CFOs to Fund Brand Marketing --- [00:17:05] Phil: And you've said that like this obsession with thinking in short term metrics is killing good marketing. And we don't always get to choose the time horizon though, when it comes to like leading marketing at startups or a bigger company. [00:17:19] Like what should marketers do when leadership demands results this quarter? Like I feel like it's, it's a bit like utopian to say that, you know, like, and, and I agree with you a hundred percent, but you know, being able to say like, we should do both. Like we should educate the market, we should do brand investments and not just demand gen and performance, but then like you meet with the CEO and like the metrics are red and the pressure is on, and it's all about short term wins. [00:17:47] And the turnover for A CMO is the shortest it's it's ever been. Right now. What pitches to CFOs don't work and what language does, when the CEO says things like impressions and blog views, like, [00:18:00] don't pay the bills, how do you, how do you counter that? [00:18:03] Guta: I love this question, and I think it's probably one of the reasons that I started, [00:18:09] like I sold the first company and started the second one. I. And I always had this, I always had good conversations with, with CFOs, with leaders. And let me just go back to one thing that you mentioned. I think we do have a problem with brand marketers right now 'cause we are having this paradigm shift. Our is being more scientific and data driven. And I think most of the brand marketers are trapped in a conceptual past [00:18:41] Phil: Mm-hmm. [00:18:42] Guta: wondering about purpose and concept. And those are important things, but you need to understand how to connect that to business results and to talk, um, the business, [00:19:00] like a business strategy. [00:19:01] And we understand how a business works work. So I think that's the first part of, you cannot address a CFO in a brand marketing language. Because they will go, like, I, I, I can't even, like, I can't guess. I can't, I don't know how to talk to you. I don't know what, like the vibe this scholar gives. I, I can't, like, I can't connect. So I think that's, that's an important part of important shift that we need to, to do in the market. And let me talk to you in a, like as you were A CFO, I'll talk to you in a p and l perspective. I think that's the only perspective that, and I get, I get cash flow. My first company, we didn't have any investment, like not [00:20:00] investors. [00:20:01] Now we have like the top VC firm in Brazil. But in the past, like it was a couple girls, couple MacBooks, that was the initial in the [00:20:09] seed investment for my first company. Boots up like glitter, [00:20:13] literally. So I get it, I get, I mean it's nice that we are talking about the long term, but we need to be alive in the long term. [00:20:24] Phil: Right. Exactly. [00:20:25] Guta: So how do we do it? Alright. So, and I think there are two ways of addressing like two business philosophies. One of them is the ones that I don't need to convince. 'cause they already believe in branding and I think they are more common in the US and Europe. Then let them, uh, and that's our philosophy here. [00:20:47] We don't, we don't call to act our clients, we don't tell them we don't like, we don't using, we don't go shop, buy and stock. But those guys already understand the [00:21:00] importance of branding. I'm talking now to the ones that doesn't. [00:21:02] Phil: Mm-hmm. [00:21:03] Guta: Uh, marketing is a cost performance marketing is a cost, right? So you need to add this cost to every product and to every, every service that you sell. So if you need to pay Google or Facebook meta or TikTok or whatever TV, broadcast, whatever to every product that you sell, then you need this portion of your margin, of your cost structure to go to marketing channels. And that's how you think about performance. How much do I need to bring this customer on every time they come buy from us? Problem is it reaches a ceiling. It will only get more expensive. That's call saturation. [00:21:52] And it's because you are ready, so smart to the most obvious customers and now you are. You need to reach [00:22:00] a more distant audience and you need to impact the same people again and again 'cause you don't know who is in market for your product. Right. Maybe the platform knows, but not really a hundred percent. So it, it'll only get more expensive, but you need to, to have it in your cost structure. The performance marketing that goes into cost, I mean economically a cost. It goes into each product. What if alongside with the cost, we had some investment, like it's a different line of budget that goes into investment, that goes into decreasing the, this cost over time. So we will invest in things that generates themselves over time. So the larger a brand is, the more sales it gets. It's been proven a times in [00:23:00] research because more purchase, more buyers generate more buyers. And a larger audience generates our audience. So that's organic, right? Uh, so what if we could create something that lasts in people's minds and that over time it'll decrease our individual cost of marketing? 'cause it will generate a baseline of customers that go, come back to us, call it branding, call it wherever. It's building stuff that decreases your cost. And it goes into the young, it's not a cost, it's a, how do you call it? An English? Like it's an investment. It's a, an expense, [00:23:49] Right. [00:23:50] It's an expense that is saying English. [00:23:52] Yeah. So that's, that's it. Uh, all right. But we don't have money to do it. All right. Would you give me 5% [00:24:00] budget to test. 5%. I mean, every CFO agrees to a 5% budget testing into anything that will bring more immediate results. I'm talking about a month, couple weeks, that's it. So you should start doing branding to things that will generate immediate returns and then you will expand your time horizon as you grow your audience. 'cause then you're reaching people that buy less frequently from your category. So yeah, so I think that's the way to think economically in a bus business way about how to invest in branding. It's not that let's do this huge campaign and spend all of our budget in crafting a message that we don't even know if it works, if it connects, connects with clients. [00:24:53] Phil: What's an example that you have in your head of like a 5% budget that's going to lead to [00:25:00] returns in like a week or two or one month? That's like a brand investment. Give us an example. [00:25:05] Guta: I can give you one that we did here. [00:25:08] Phil: Yeah, sure. [00:25:10] Guta: So we started the TikTok. We were the first, first B2B TikTok, big account in Brazil, like big viral reach. Uh, the TikTok team used us as a use case multiple times, and so we had no budget 'cause we are a tiny startup in Brazil. I looked, I, it's always about looking at the audience, at your clients. So we had this WhatsApp group. We always built community, which was always part of our marketing strategy. And they went on and on about using TikTok. It was a couple years ago. 3, 2, 2 years ago. Yeah. So, and people, marketers were like, Hmm, are you guys on TikTok? [00:26:00] Because I need to do to run this campaign, but it's all about kids dancing and I don't really feel it. And I was like, they're all bored on TikTok during work hours. It's not a social network like to them. 'cause they don't have friends on TikTok and they don't get interesting content in Portuguese. [00:26:28] Phil: Mm. [00:26:28] Guta: So let me make videos about branding and I started talking about how, how to drive results using branding. Now we have over a million views of million likes and almost a hundred thousand followers and am a link in Top Voice and so on. [00:26:46] So it grew bigger, but by the time. We started doing videos on TikTok on my spare time and listening to the listening, like reading the comments and talking to people and [00:27:00] seeing how they would react to the content and make more of it. And then it grew from there. So it wasn't really a 5% budget, it was a 5% energy, but let's call it like that. But I think it's like that you need to observe how your client is buying and where they are at, and what they value about you and about your brand. And then do this one channel, one thing, see how people respond, and then expand from there, and then pick a second one and so on. [00:27:36] Phil: Very cool. I, I love the way that you break this down for A CFO instead of saying, we need a 5% budget to do branding, and we need branding involved in this because it's mindshare, it's perception it's gonna help sales in the long run, if not like in the medium term, you're kind of flipping that script a little bit and saying, instead of calling it branding, let's just say [00:28:00] there's this thing we can do that's gonna reduce the cost of acquisition. [00:28:03] It's gonna reduce the cost of marketing. In the medium term and it can even happen in the short term. And I don't know, for, for A CFO, I feel like that's an interesting way to say it. 'cause as soon as you say branding, the perception is long-term, hard to measure, won't see this in like six months, a year from now. [00:28:23] So I like how you're saying this thing that we're gonna do, we don't need to call it branding, but it's gonna reduce cac and in it, it might have immediate returns on that, but like the TikTok example that you just gave, like were you seeing like results within one week of that? Like what was the CFO's first reaction like when, when you first did this? [00:28:44] Guta: yeah. Uh, in, we, we track through our product, we can see the time lag of, of returns and some brand actions do, uh, happen immediately. Do have immediate, uh, or one [00:29:00] week conversion time lag. But to this TikTok, I think it went viral like very early on, probably the second video. And then our sales rep started telling us that the leads were watching, and then we started having more leads. I think that's TikTok. 'cause that's the only thing that we did. I don't recall if it was like immediate, immediate, but it was like before I gave up. [00:29:36] Making, [00:29:38] Phil: That, that's always a strong signal. When the sales team says that, it starts showing up in sales conversations, I. [00:29:44] Guta: It's qualitative branding, right? [00:29:46] Uh, because before you can really measure it, before you have a product like ours, uh, it's reading. They come, it's a bit of feeling. And that's what at the same time, if you talk about numbers with the branding [00:30:00] person, there's their brain shuts [00:30:02] Phil: Mm-hmm. [00:30:03] Guta: And if you talk about concept and branch with your fold, their brand shuts down. So that's the difficulty of this, of this conversation. [00:30:12] And, but it, but they do, like, they do know deep down. 'cause I've seen CFOs go on and tell like, why don't we hire this influencer? They know it works, right? And sometimes they go, oh, I've listened to this podcast, it's humans of MarTech and we should definitely sponsor that. Right? So they do like, they do think of branding actions sometimes what? Like they get it. The problem is it's very hard to bet on something that you don't have data to back it up. And if you are only doing one thing, it's easier. But if you are somehow mid-sized to large company, you are doing all these things at the same time. And [00:31:00] how would you commit to something if you can't measure? So that's, that's why I think we do need like a unified measurement that's part of our pivot, uh, from research. 'cause we were measuring only branding. And I think the changing marketing needs a like unified model where, where you can see everything. [00:31:21] ​ [00:33:24] Phil: Maybe, maybe give us another example. I'm, I'm curious like, uh, because [00:33:28] 5. How Attribution Models Handle Channel Composition and Forecast Errors --- [00:33:28] Phil: you're sitting on a lot of data from customers using the product now, and like you said, it's not just branding stuff, but you get to see and tell customers what is working and, and what is driving revenue by looking at this data. [00:33:41] Like what are some examples that come to mind that were maybe like a bit surprising to you, or you think in the market people assume they are undervalued or underrated and not enough people do those things. What are examples that come to mind? [00:33:55] Guta: I like, I like the examples that are not the same [00:34:00] within clients. So for example, influencers, they work for some clients and don't [00:34:05] work for other clients. So that's interesting. And pr. PR is something very difficult to track. Not 'cause it does, doesn't work, but because we don't have real numbers. I dunno how many people have seen my article that I published. The largest business magazine in Brazil. I dunno how many, what will I add to the model in terms of data? So that's very difficult. Uh, but we do see, um, TikTok awareness campaigns generating results or generating other things. Like that's what I like the most when I see TikTok generates blog views. 'cause then you try to understand the consumer journey somehow, right? Oh, they discovered there and they, they went [00:35:00] on to this and then what? They started a free trial and what I, there's a client of ours that's spending less than selling more. [00:35:11] Phil: Hmm. [00:35:12] Guta: So, 'cause it's about the ament of things as you do. It's not 'cause people like this, this idea of a magic channel. Right, because it's easy, but it's actually, uh, the composition. So I think that's another insight. It's about the, the composition and then there's another that just happened. Um, there is a client of ours that, so TV broadcasts in Brazil are huge. Um, so we have every day the audience of the Super Bowl. That's how big, like every day there is a Super Bowl on TV here, in terms of audience. It's crazy. And so they did TV and it worked, and they decided to [00:36:00] do this again, a new flight. And they did the investment and the algorithm forecasted like a huge sale and like peak. But then it was wrong. Like they didn't sell. [00:36:15] They were like, oh, even shit. Even the algorithm thought that we would, um, sell. And we didn't. So they went back to analyze creative and what else was happening. So it's always, it's also about, okay, it should have sold from, like you, you made the right call, but somehow something changed. So that's also something about forecasting. That's that's it's a bit, a little bit, well it doesn't always happen as forecasted 'cause things change. [00:36:48] Phil: Mm [00:36:49] Guta: And there is another case which like recent one as well, we saw this huge increasing baseline and attributed sales conversions and [00:37:00] what, what's wrong with the model? Like is the algorithm wrong? And then we went to talk to the client, like ready to listen a lot, and they were like, oh, it's about the new legislation. [00:37:13] Phil: hmm. [00:37:14] Guta: And we were like, oh, really? Yeah, because there was something that they couldn't communicate. Because of some, some change in the legislation. Uh, but they, they had the big, like a big player lead sales to them. 'cause they became a partner in this new structure. So they, they, they, got a lot of clients from this business partnership. [00:37:39] They, they could not communicate. So it wasn't as rebutted to any marketing channel. And that's true 'cause they couldn't communicate. Uh, but it did increase in sales, so they had a pick. So it's interesting to see that, that those type of, uh, like early technology things like, I remember like in the beginning of the [00:38:00] iPhone that you would like shift it and these things would move. [00:38:03] And we were like, oh God, look at it. I think of discovery. [00:38:09] Phil: Very cool. I appreciate you sharing that. So let, let's get into the product a little bit. 'cause like, if, if folks are listening and thinking like, alright, it sounds really cool. It's amazing that you can get this data. How are you actually getting this data? [00:38:22] 6. How to Build Trust in Marketing Measurement Solutions --- [00:38:22] Phil: So the product that you launched in the summer of last year in 2024, you're attributing marketing efforts and you're showing actual impact on business results. [00:38:31] You're including branding actions in there that usually get left out of a lot of click-based, uh, attribution models that we have today. So you're including things like influencers, pr, uh, out of home media, sponsorship, podcasts, organic content. You do this by using statistical modeling on top of data the company already owns. [00:38:50] But we were chatting before we press record. Like I feel. We're in this like dark age of marketing measurement. A previous guest said this on the show where like, marketing leaders [00:39:00] don't know what to trust. Like is it m, MM that I should do? Is it incrementality? Is it MTA? Is it causal attribution? Some experts claim that like, this is the thing that you should be doing, and others are saying that's terrible and here's what's wrong about it. [00:39:14] And like the things are changing all the time. A lot of strong opinions about it. Some people say that causality is actually impossible to do in marketing. We only have correlation. How do you build trust with buyers that your solution gives truthful data? When you say actual impact, like how do we know it's actual, what are your thoughts there? [00:39:36] Guta: I love this time. What a time right? To be on marketing. [00:39:39] Phil: It's true. [00:39:41] Guta: I do love it 'cause it's a shift of paradigm in two ways. We're going through a shift in paradigm in terms of, um, data driven approach to marketing, in terms of mentality, ethos. But we, we are always also [00:40:00] going through a shift in technology. Through lms, right? So there are two revolutions going on in market the same time. And then there's a cultural one as well. Post COVID. We are working differently, we are more global and all those thanks. Alright, so how do we build trust? That was your question. Not about the the environment. Um, but I do believe that, and I'm leaving both changes of paradigm at the same time in tech and in marketing. And I think the answer to both of them is community. And I don't think that someone alone can change an entire industry. I don't believe in a genius cracking the code. Uh, it took us four years to reach this product. Why? 'cause on the first product that was branding research that. We read [00:41:00] all those academic papers, I can't even recall, like over 30, 40. And I talked to all those, I mean academics, my family, super academic. Um, and then we went, went on to, to produce content. And I talked, when we started this pro, when we decided to pivot, I posted like, can you share with me your metrics, uh, spreadsheet. And I got like, obviously blank, but the structure of the spreadsheets of like the top 10 companies in, because they shared with me our clients. [00:41:38] They share links like let me show you how we are AB testing. Let me show, uh, we are officially launching 'cause we pre-launch the product with 10 companies using to train the algorithm. And then on Friday we'll launch officially, anyone can buy purple metrics from Friday. So now when people are listening, they can already [00:42:00] bite. [00:42:00] But I don't believe in building it alone. I think it's about the conversations. And that's why I think we are doing such a great job on talking to all these people building 'cause all these opinions are important. Why does someone, why they can't see causation? We see Caucasian, uh, in our, but it's a different statistic modeling and probably won't fit like the lastly, uh, measurement. [00:42:28] I know the context in which they thought it because I, I did listen to this, which is a great one. And so I think this conversation is very important and I think it's about creating this movement on, we need to measure, we need to be more scientific. But it's not about, let's last click. It's not about only generating sales. [00:42:52] 'cause that's a. That's generating anxiety and it reaches a silly, uh, we need to build [00:43:00] a different type of relation with our clients. And I think markets evolve to that. If you, if you think about economics, uh, behavioral economics, it's a little bit adding data to psychology. And then if you think of every, every discipline, it evolves like that. And it's about movement and it's about benchmarking, it's about, and it's about testing. So when a client of ours looks at it and says, well, yeah, 'cause there is a change legislation and it showed here, and it, you have like part, the part that's trackable makes a lot of sense with what we are seeing in MTAs or in click-based measurement. [00:43:46] That makes sense. That makes sense. Uh, like the forecasts are the same, so it's. It's about that and evolving the algorithm with data always, always a good idea. And I think [00:44:00] being aware of academic research, I think marketing is very distant from academia compared to other disciplines. Um, and I think we need to close the gap. [00:44:15] Phil: Very cool. Yeah. Maybe we can chat about an example. [00:44:18] 7. Creating Attribution Algorithms That Respect Brand Complexity --- [00:44:18] Phil: So like purple metrics customers can see for example if like the LinkedIn post from their founder drove more Google searches, which then drove lead conversions or, or sales. And you're essentially giving data on like how long this ripple effect takes to happen for customers. [00:44:39] To me like this feels really hard to do. Like a lot of folks claim to be able to do this and like seeing the relationship between a post and web website, website traffic, it is pretty clear. Even though LinkedIn doesn't have like links all the time in the post, you can still see source traffic. But modeling the delay [00:45:00] feels nearly impossible, like especially in B2B. [00:45:04] Are folks just better off asking where did you hear about us on a form? Are you also including qualitative data in in the data that you're sharing with customers? Maybe chat about that a bit, [00:45:14] Guta: Uh, well, we shifted from that product, right? So [00:45:17] Phil: right? [00:45:18] Guta: I have a strong opinion. Uh, but, but, um, yeah, you can always ask people as a, as a way to fact check, let's say. But when you ask someone, they, you'll see what has the higher memory structure in their head, and sometimes they just mark and they think. I can't. I know. [00:45:43] It's just so much. Right? And you go like, yeah, whatever [00:45:45] my friend [00:45:46] Phil: First answer. Done. [00:45:49] Guta: Yeah. It sounds like, uh, just one thing. We, we don't see how this LinkedIn post generated sales, we see how LinkedIn reach [00:46:00] from a specific week generated something else. [00:46:04] Phil: Okay. [00:46:05] Guta: You can't attribute to any specific because you need to aggregate the data. [00:46:10] Otherwise you would have too many variables. Uh, so why not just like, why don't do the, the easy path. And I think the answer is 'cause people can't run a business on, on declared answers. That's why we should, I mean, it would be easier if we could trust like your brand perception matched with your brand, [00:46:36] Phil: Mm-hmm. [00:46:37] Guta: with your declared behavior that. That's what we were doing. And I think it's easier and I think it's valid. We did. I mean, it would become a nice business. Not a huge one, but a nice business. It's a product. People were using the product. But I think your, your question is more about, it's really hard [00:47:00] and I think no one did it 'cause no one really tried to, 'cause they think we have so much to unpack in terms of measurement. We need to solve clicks in aless area ever. We have this customer data platforms challenge as well. Uh, we do need to fix creatives in a, in a, an AI era. So we have all these problems in marketing that needs to be solved. And I don't think that there's someone from branding that really gets branding. They try to do something about. Like a data driven approach to it. So I I, it took me like almost a year to start this company. 'cause I was like, someone might have solved it already. Right? Because it's, I mean, it's a big problem, [00:48:00] but I think, I dunno why, I think people that tried, they tried to do brand fracking in a more, in a more technological way and using ai. There are a lot of nice solutions and I do think there are a lot of proposed solutions to all these measurement problems. And I think it's about philosophy. In the end, you need to understand how those people think and what 'cause algorithms, they do have opinions it. The next post that you see on your feed. It's opinionated. [00:48:39] Phil: Mm-hmm. [00:48:41] Guta: Uh, so we do make decisions every day about the algorithm and our decisions are made from a company that believes in building together with its clients, putting branding into the equation, and creating a unified system to [00:49:00] marketing so you can manage, like having a lower cost over time and a more solid baseline in organic channels. So this is what we believe in, and if you don't believe in it, if you believe in a more, like in the media buying that's more effective within like the refinement of the media channels, we are probably not the best supplier. So I think marketers will start to think on what do I believe in and, and look for. Provider, an algorithm that's built for it. Because I mean, you have open sources, you have like, you can go for it. Like, so, so it's, it's, still early days of this market and I think time will tell and when we see the client, like comfort confirming what they are seeing on the algorithm. And it [00:50:00] makes me a bit more, I remember on, on early days, I was like, I went to our like, lead data, scientist lead. [00:50:08] I was like, would you, like, I know the algorithm is saying to do that, but would you do it if you were the CMO [00:50:16] you would you, what would you do? And he answered, no, I wouldn't. I was [00:50:22] like, fuck, [00:50:24] Phil: Shit. [00:50:25] Guta: shit, what happened? And then he, he went on to say, look, I believe the algorithm is 100% right. But I think you might reach the same with less shift in marketing, like [00:50:42] the same, like very similar result, changing a little bit less, but then you do, you wouldn't have to, 'cause when like it's easy for an algorithm to say, take it from Google and put it on TikTok or on out of home like billboards. It's a different [00:51:00] team, [00:51:00] it's a different agency, [00:51:02] it's a different, everything, [00:51:03] like a different campaign. So it's not that simple [00:51:06] to shift money around. It's not like my investment banking, whatever, that I can just buy Bitcoin, you know? [00:51:13] It's not like that. So, uh, and that's when we, we developed something that we call baby steps. So there is a safety lock in the algorithm that doesn't show the entire shift. It's proposed [00:51:30] Phil: Hmm. [00:51:31] Guta: and it shows like part of it. So the marketer can shift like money and then the algorithm will learn and improve the recommendation for the next week. So I think it's a process of evolvement and building together. [00:51:46] Like that's something that we ask, like I asked him, but they, we could ask our clients and we see how they react to things and we input that into the algorithm. So we'll get better over time and I think the entire market will get [00:52:00] better. [00:52:01] Phil: Very cool. Uh, I'm, I'm curious to ask you like, [00:52:04] 8. How Algorithm Blindness Limits Predictive Marketing --- [00:52:04] Phil: how white box is the algorithm? Is it like a black box? Um, maybe you're thinking about this already, but like, I think the future of this space is allowing your clients to have a conversation with the algorithm. When it suggests something, you get to have a conversation like, all right, why, like, unpack why you're suggesting we shift everything to out of home. [00:52:28] Show me the data points behind that record. And if I give you a bit more context on the legislation that just came out locally on that might have triggered that thing, or like there's this economic shift that's maybe like Trump in the US said this on true social, like giving it a bit more data and then seeing, like having that conversation going back and forth. [00:52:50] Maybe you're thinking about that already, but like. [00:52:52] Guta: I am, I am. Uh, it's interesting your point of view and I am in a slightly different way. I think we do [00:53:00] what will, we will have more conversational front end ux, uh, to, to, and I do, we do it right. So when I see some, like this baseline pick, what happened is that, uh, our CS team went to talk to the engineers and they were like, what the fuck? And so, and the data scientists went back and said, look, the data is correct. And then they were like, all right, let's talk to the client and see if it makes sense. So I think this, uh, what's going on? And the algorithm will probably answer, look, the data is correct. I mean, you don't have missing data and it's consistent and it's pro like, it'll provide you with some answer. [00:53:54] But I think something that [00:54:00] annoys me is that the algorithm is blind. It doesn't know what goes on Instagram. It only knows it had more reach or more likes. And then that it probably sold more, brought more leads on. So I'm thinking about ways to inputting this data into a, the algorithm, but I think it needs to be in a more systematic way. [00:54:27] I don't think it will be fuel telling the [00:54:30] algorithm that Trump said something. [00:54:32] Phil: Yeah. [00:54:32] Guta: Uh, I think it'll be more like what happened there in a more qualitative way, but doing it automatically. So that's what I'm testing in our lab right now. Um, so I would love to have like this new version of the product that's more qualitative in terms of insights and more conversation. [00:54:56] But I think like in a year maybe, I don't think it's a [00:55:00] short term thing. Um, and I do think that sometimes, you know, Trump will say things and it'll shift the economy and the war will start and it'll shift the economy. And if you think about it, when I use any prediction algorithm that happens, right? So if I get into a car for a road trip, like if I go to the next neighborhood, it'll probably be very accurate ways or Google maps. Uh, but if I'm going to a different city, if I'm driving a hundred miles, couple hundred miles, it'll probably be somehow wrong. But I'm okay with that. I know that's the prediction from the time I went out and that's what I wanted. And I understand if something happened on the road on the way, and when I'm packing to go [00:56:00] to a different city in the different country, I know that that temperature won't be precisely, that I might put in a, an extra jacket or, you know, bring in the swimsuit. [00:56:10] 'cause just in case so. We are used to predicting those type of things in our daily lives. And I think we'll get more used to it in marketing. So we'll probably get more used to, uh, you know, that was the forecast, but we messed up with something or something nice happened on the, along the way. This, this, we, we have a way in the algorithm to see if something's trending, which means the algorithm forecasted fewer sales from, its somehow how it's selling more. 'cause you did something, go look at it. [00:56:46] I don't know. Go look at it. The prediction is different from reality and in like in a suitcase, you don't really need to know about the cold masses that generated the [00:57:00] temperature drop. But maybe on the road you, you do want to know that there was an accident. [00:57:04] Um, in marketing, you do want to know, I. What happened that generated the different outcome from the predicted one? So I think it's part of the ment of the, of the market. Yeah. But I think we'll be in a very different place in a few years. [00:57:20] Phil: Very cool. It's, it's exciting and yeah, really interesting to hear your, your point of view on that as you're, you're building and testing stuff out in the lab with, with stuff that's already out. Um, I got two last questions for you and we're running up on time. Um. [00:57:33] Guta: So I'll be fast. [00:57:34] 9. When Marketing Data Refuses to Flatter Your Strategy --- [00:57:34] Phil: This one topic with attribution, like, I feel like attribution gets like a really bad rep because oftentimes A CMO will buy the attribution tool to justify their existence, and they will say that they wanna like invest here and invest there. [00:57:50] And then they buy the MT A tool. And the MDA tool allows them to like tweak the levers. Like, oh, we're gonna DW shaped instead of U-shaped because my [00:58:00] brand investment looks way better. There's way more credit given it to it there. So like what are you, like, how do you think marketers should react when their data challenges the, the data that you show challenges their instincts? [00:58:12] And how are you building purple to service those uncomfortable truths without turning it into like a self-congratulatory, uh, performance tool that just like validates everything you're already doing. [00:58:23] Guta: Uh, that's a great question and I think it's my answer goes to actually to the company leadership. 'cause I think it's about company culture. I do see, like, I have clients that rather churn, like they prefer churning, like canceling purple metrics than facing the fact that they are branding is not working. Uh, yeah. So that's that's true. And the more a company gets political, the more like those things starts happening and I think we need to sponsor [00:59:00] cultures when it's okay. To fail and learn. So if a CEO cannot get to the board and said, we did it, we have a client. And they, they, they were laughing 'cause we were like, what's did this about? [00:59:16] And they were like, oh yeah, we messed up. Learned about it next. You know, so, so that's a great culture. Uh, and I think you need to be ready to have data. So most CEOs will tell you, most board members will tell you they want data. But if they don't want, that means if you want to be more scientific, science is about learning by experimentation, which means failing and not failing and learning from it. That's all about that. That that's what it's all about. That. So if A-C-E-O-A board member wants to be more data driven, it should mean there should be room for failure to. [01:00:00] Within the company. And if A CMO is concealing their mistakes, it's probably 'cause they're afraid to be, they will be fired if they show this. So it's about we need to start to having this, when someone says, where's the data? Like, can you handle being wrong? Us being wrong, like, we'll do everything to be right, but we will fail sometimes. So we need to be open. I need to be able to be open about it. Can I, let's agree on the strategy. Let's make sure when we make the decisions, we were backed by data and we were sure, we were certain that this was the best test, the [01:00:45] best we could do right now, uh, with the information that we do have right now. [01:00:49] But in the future we'll have new information. We cannot penalize our like. Today, [01:00:56] uh, our present selves on information that we'll get in the [01:01:00] future. So I think, um, that's a cultural shift [01:01:03] and I think some companies will get there faster than others. [01:01:08] Phil: that's such a good point. It's like almost having psychological safety in that culture of data and knowing that. If we try something and it fails, like it's okay, and we're not gonna cut funding to marketing just because there's one failed experiment. As long as we're all aligned on this is the way that we're measuring this and, and we're bought in there. [01:01:29] Love, love your answer there, Guta. [01:01:31] 10. Building a Personal Happiness System Like a Product --- [01:01:31] Phil: I got one last question for you. Um, you're obviously a CEO. You're known as the queen of branding in Brazil, but you're also an avid urban life traveler. Uh, you're big time in sports and you also are a bookworm. You got a ton of stuff going on in your life. One question we ask everyone on the show is how do you remain happy and successful in your career, and how do you find balance between all the stuff you're working on while staying happy? [01:01:54] Guta: I love this. That's the final, the final question, and I learned so much from other people's, uh, [01:02:00] happiness systems. Uh, as a designer slash branding slash product person, I, I approach it like a product and I think it's something that you need to build up from on top of, like you need to start and then build features on top of features. [01:02:20] Phil: Hmm. [01:02:21] Guta: And so now I'm very aware of the things that may make me feel good, which is sports in, in my routine. Uh, I need to. Be around water. Like once every, something. I'm from Rio. I grew up in Rio. Uh, we have like the beach, the sea. I need to go back to the ocean. Every once in a while. I need to be around my friends. I need to be reading. I love disposing movement of the city. Uh, so I structure my routine in a way that I make sure [01:03:00] I'll have time with my friends. I'll be traveling every once in a while. Um, I'll be like, you know, all these things and I need to have this intellectual challenge. So talking to people like you or even listening to podcasts or reading, so I understand the function, uh, quote unquote of this stuff that makes me happy and try to add it to the system. And I also have this. Like a bug report when something's [01:03:36] wrong. Like when, why am I so tired? Why am I so stressed? So I try to understand those signals, like those metrics on unhappiness. And then I unpack why am I so stressed out? What's missing? Oh, I'm not, I'm not hitting right. I'm not sleeping well. I'm not, so I go back to fixing the system. [01:03:57] So that's, that's my approach. But, [01:04:00] and it's a constant, you, you can't, like, you need to dedicate time to it, otherwise you just got, get run over by work or routine. And then, yeah, [01:04:13] Phil: Still have a, a product roadmap for it. [01:04:16] Guta: I, I do have a little bit [01:04:19] Phil: Nice. [01:04:20] Guta: things that I wanna do, but I try not to be so long, like, 'cause the long term generates anxiety. [01:04:29] So I have a list of books that I wanna read. [01:04:31] I. And I always know where I am traveling next. Uh, you know, those type of [01:04:38] like small [01:04:40] Phil: very cool. Love the analogy there, and no one's ever said that on the show. [01:04:44] Guta: really. [01:04:44] Phil: yeah, yeah. Think of your, your happiness and, and balance in life as a product. I think it, it makes sense for you, obviously building a product and tracking the effectiveness of other companies marketing. So yeah, love, love your answer does super fun. [01:04:58] Uh, we'll link out [01:05:00] to, uh, purple Metrics and the blog. Uh, I think, uh, the TikTok also, uh, I've also [01:05:05] Guta: in Portuguese though. But I did record a few videos in English, [01:05:09] uh, recently. So [01:05:10] Phil: and, and the LinkedIn stuff that you repost too, like LinkedIn has like native, uh, translation in there. So, uh, I was able to catch up with a, a few ones. I think what you're doing is, is really, really cool and I really appreciate you, uh, sitting down with me for an hour. Thank you so much for your time. [01:05:24] Guta: Thanks. Same here. Thank you so much.